Multiple-reader environments with shared filesystem?

Hi,
we are thinking about offloading some read-only-tasks to a different server so that one server opens the JE-Environment read-write and the other server opens it read-only. What kind of filesystem is recommended to support this architecture?
The easiest solution is of course using NFS to export the BDB-Files to the read-only-Server. But what performance does one expect?
What would one do if there is a SAN storage infrastructure already available? Could one use Oracles OCFS2 to share the filesystem as a block device at SAN-Level between the two servers? One critical issue is that "flock" is not supported by OCFS2. Are there any other Filesystems that support multiple readers (multiple writers are not needed!) on top of a SAN?
Any other ideas/suggestions for a high performance solution for a shared BDB-Environment within a SAN storage infrastructure?
Stefan.

Hi Stefan,
Just to be clear, I am assuming that you do not want to access the r/w Environment via (say) NFS, but you do want the r/o environments to be over (say) NFS.
We require flock to work properly so that would rule out OCFS2 (I confess ignorance on OCFS2 so I'll take your word for it that it doesn't support flock). As long as your NFS implementation supports flock, then NFS should be fine -- if it doesn't then bad things will happen. Performance over NFS will be based on how big the file system cache is on the r/o side, etc. I can't think of any reason why performance would be bad, but you should obviously try it and see how it is (please let us know).
Regards,
Charles Lamb

Similar Messages

Debug Multiple Flex Projects with shared PHP project

Hi ! I have a problem debugging Flex/PHP applications.
I have many Flex projects with a shared PHP project, so I need to set the the output folder to the Flex project's folder, not the PHP's project folder.
But it doesn't work, I mean the application works great, the PHP are executed fine, the AS3 code is debugged fine, but I can't debug PHP.
If I put the output folder to the PHP's project, then I can debug the PHP code too.
Please, can you tell me if it's possible to put the output folder to another folder than PHP's project???

Hi Steven,
Yes, I considered this option too.
But if someone creates in the database a column for a table that belongs to "Project A", when I publish "Project B" the generated script tries to remove the column since my copy of "Project A" on my solution is not the latest
version.
I think I'll just have to handle this. Like, organizing with people when and ho can publish to the databases.
Maybe the best practice is to developers use their own LocalDB while developping and only one person publish to database on the development server.

Multiple product install with shared configuration

Maybe what we try to achieve is simply not possible. I can get it to work if I make manual modifications but I fear that the resulting setup will fail after an update to the p2 profile (update of a IU or new IU..)
The setup we want to achieve is as following:
--Installation folder, can be written to by administrator (on windows this could be program files for instance), he has p2 tooling that he can use to make modifications to the setup, these tooling will request elevation so it all plays nice.
--Executable come in root
/InstallFolder/Application1.exe
/InstallFolder/Application1.ini
/InstallFolder/Application2.exe
/InstallFolder/Application2.ini
--Configuration folders to which osgi.shared.configuration.area should point in the ini
/InstallFolder/configuration/application1/config.ini
/InstallFolder/configuration/application2/config.ini
--Shared P2 agent with bundle pooling
/InstallFolder/p2/org.eclipse.equinox.p2.engine/profileRegistry/application1.profile
/InstallFolder/p2/org.eclipse.equinox.p2.engine/profileRegistry/application2.profile
--Bundle Poll
/InstallFolder/artifacts/plugins
/InstallFolder/artifacts/features
The osgi.configuration.area should point to a folder under user.home which we assume is always writable.
The problem is that if we try to enforce this setup by setting the custom configuration folder in the p2 profile (org.eclipse.equinox.p2.configurationFolder) it will always write this folder in the ini file as -configuration {folder}. As a result the launcher ignores osgi.configuration.area in the config.ini or launcher.ini which points to the userhome.
I cannot fill in the user.home in the p2 profile because than other users will not have a working setup. Am I missing something here?
I could manually remove the -configuration after our installer and then the application will work but since the equinox manipulator always seems to load and write this -configuration {folder} if it is not 'configuration' I fear that this will break the setup after one touches the profile to install extra IU's which most of our customers do.
Our current work around is to have each executable in a folder and then the default 'configuration' folder can be used. This feels more like a hack but it works because the -configuration is not written in this case because is it 'configuration'.

schneida wrote:Are there easier mehtods to have a ArchLinux install with local packages?
Personally I use a squid cache proxy to keep a copy of the packages that will be used across the various computers.
The interesting sections of my /etc/squid/squid.conf file are;
maximum_object_size 500 MB
refresh_pattern abs.tar.gz$ 0 20% 4320 refresh-ims
refresh_pattern db.tar.gz$ 0 20% 4320 refresh-ims
refresh_pattern files.tar.gz$ 0 20% 4320 refresh-ims
refresh_pattern pkg.tar.gz$ 10080 90% 43200 override-expire ignore-no-cache ignore-no-store
The patterns will
* Cache all pacman packages for at least a week.
* Check the server if the database has been modified since it was last accessed.

Multiple hierarchy levels with shared parent

Hi All,
I have multiple hierarchies having same parent level for a dimension. For example for a given dimension we have two multiple hierarchies
A->B->C->D->E
A->B->F
How should we set up the dimension hierarchies so that we can drill through both the hierarchy path having the option popping up for the user to decide?
Thanks,
Virat

Yes, I have clicked the ragged hierarchy option. I can create only one dimension hierarchy of a single dimension. I want to create multiple path on the hierarchy with same shared parent level. I could see that option while creating the hierarchy,however, on the report , only one path is working .There are numerous paths for the hierarchy in the requirement and I cannot create separate dimension for each path.
Edited by: 872073 on Oct 29, 2012 7:17 AM

With Multiple Management Services, How BIg Is Your Shared Filesystem?

I am installing Grid Control 11.1 with multiple OMSs. Oracle uses a shared filesystem in this configuration for load balancing and fault tolerance. I want to know from anyone who has installed multiple OMSs: how much storage does Grid Contorl use in the shared file system in your implementation. Please also state how many targets are monitored. Thanks.

The size of the shared location is dependent upon:
1. How quickly agent files are processed - If your OMS are going to be down for awhile, size the location to support storing the files until processing can being and catch-up. The files are small but depending on the number of targets this can add up.
2. Usage of provisioning - Ensure you size to support all the various files used in your provisioning
We have several hundred targets, two OMS, and make limited usage of provisioning (software and agent installation images). We allocated an NFS share of 50GB but typical usage is <20GB, with 95% of that is due to the provisioning files.

How to config a shared cache for multiple environments with C API

How to config a shared cache for multiple environments with C API? Just like Java edition. Chapter 2. Database Environments
I want to open large number of databases, at least 10,000. But as the counts of databases opened increase, the db->open operation become very slow. It almost cost 2 hours to 10,000 databases.
So I try to distribute these databases to multiple environments ( for example, 5 envs ). And in order to improve the efficient of memory use, I want to share cache between envs.

Hi,
It is not clear what you meaning about multiple environments. Do you mean these environments are in different directories or in the same directory ? If you mean environments in different dirs share the same cache, it is interesting why you need that.
If you do not use DB_PRIVATE to open the environment, the created cache will be on disk, in the environment directory, so it can be shared by multiple processes and multiple threads. Currently, the cache file is in the environment directory, and we do not support specifying a separate directory for cache only.
Regards,
Oracle Berkeley DB.

Are there suggestions for managing Multiple VMWare Fusion systems sharing files with a MacBook? Suppose I have enough ram for 3 servers to share files.

Suppose I have enough ram for 2 or three VMWare fusion systems to that could benefit from file sharing. Is it possible to configiure each VMWare funsion instance to have a common read only file share on my desktop and a common read/write area on on my desktop?
This would allow me to extend the power and mustle of my Mac Book Air, Retina Display system to its best advantage. Whether native Mac processing or speciality linux or Windows systems as needed.
I am having trouble allowing more than one system to share the same read only and the same read/write areas so that software does the file locking and merge controls. Even if these do not share a common read/write area, it would be truly handy for drive space management for them to share a common read only area.
Thanks for your thoughtful input.
20 year IT Professional that loves Mac visual interfaces, outstanding performance and a *nix command line.
Don

It really denpends on the os running in the vm and what the permissions are set to on those shares.
I run a few vms on both a mac at home and windows at work and all other computers in my home or at work can connect to the vms without any problems. Even multiple computer at the same time.
It also depends on how you have set the vm up for networking. If you are setting it up as nat, instead of bridged, that can also cause problems access it from multiple system. With nat the vm is on its own network branch.

Is it possible to access a single Lightroom catalogue mounted on a shared drive from multiple computers, each with an individual Lightroom install?

Is it possible to access a single Lightroom catalogue mounted on a shared drive from multiple computers, each with an individual Lightroom install?
Thanks.
Steve Greenberg

No, Lightroom catalogs cannot be on a networked (shared) drive, and they do not support simultaneous multi-user access

Read MULTIPLE idocs data with all sgmn to my internal table in a single

Dear SAP Folks!
I have a question, I know to read SINGLE idoc data with all segments, we have FM IDOC_READ_COMPLETELY but my requirement is to Read MULTIPLE idocs data with all segments to my internal table in a single shot without keeping this FM in loop! since it is performance issue, at a time i may want to read 1000 idocs data!
Could anyone please let me know is there any way to get this, keeping mind of performance and any other FM?
Best Regards,
Srini

Hi,
I know idoc numbers and i can write a select Query to DB table EDID4 to get all segments data to my internal table. But i am looking for FM to do this. because i have too many number of idocs and each idoc is having many segments(I am thinking in performance point of view!) The FM IDOC_READ_COMPLETELY can accept only ONE IDOC number as import parameters at a time to get segment data. in similar way do we have any other FM? or other way? except select query to EDID4 table?
Best Regards,
Srini

Announcing the availability of enabling Windows Server 2012 R2 Essentials' integration of Microsoft online services in environments with multiple domain controllers

In Windows Server Essentials 2012 R2, all of our online services integration features, including Azure Active Directory and Office 365, are supported only in environments that
have a single domain controller. In environments with more than one domain controller, integration of these services is blocked due limitations in the user account and password synchronization mechanism in Windows Server Essentials.
I am happy to announce that with the recent Windows August Update released on (8/12/2014, PST), this limitation has been removed. This update adds support for both Azure
Active Directory integration and Office 365 integration features in domain environments consisting of a single domain controller, multiple domain controllers, or Windows Server Essentials as a domain member server.
For more information, please go to
http://support.microsoft.com/kb/2974308

Hi JoeBeck,
Thanks for the comment. Could you please tell which link you clicked to download?
Please go to PinPoint check details and start download
http://pinpoint.microsoft.com/en-US/applications/Dynamics-CRM-Online-Add-in-12884966386
Thanks,
Shanghai Wicresoft

Read Statement with Single key with multiple values

Dear Friends,
In ECC 4.6c, I wrote statement like this.
Read table message with key type = 'E' or 'S'.
if sy-subrc = 0.
Populated the message log to output internal table.
endif.
But when i transported to QA which is ECC 6.0, it is giving error.
Can any one help me in modifying this qurey.
Plse read my question thoroughly and reply me back.
Regards,
Santosh Kumar M

Hi,
You can't use OR condition with READ table. it is always AND condition.
Here 2 ways for ur problem.
If u want to have just one message in the log then use multiple reads.
READ TABLE message WITH KEY type = 'E'.
If sy-subrc IS INITIAL.
populate log.
ELSE.
READ TABLE message WITH KEY type = 'S'.
If sy-subrc IS INITIAL.
populate log.
ENDIF.
ENDIF.
If u want all messages then u have to loop through and populate the log as sugested by Vijay but
Without exit in loop.(Since u need all the messages)
LOOP AT message into wa WHERE type = 'E' or type = 'S'.
populate log.
ENDLOOP.
Thanks,
Vinod.

How to read multiple arguments separated with space in one line

How to modify the
public static void main (String[] args){
}so that it can read multiple arguments separated with space in a single line?
e.g.
java myprogram username password host
java myprogram2 ipaddress port
Thx.

public static void main (String[] args){
int index = 0;
for(String arg : args) {
System.out.println("args["+(index++)+" = "+arg);
}

SSIS project - read multiple flat files with different formats

hi all,
i need to import multiple flat files with different formats into different tables of the sql server database and not able to figure out the best way out in ssis to do so...
please advise the possible methods in ssis to do so and if possible the process which can be dynamic as file names or columns might change in future.

Hi AK1987,
To import flat files with dynamic columns, we can use Script Task inside a Foreach Loop Container to parse the first row of the flat file to get the columns names and save them into a .NET variable, then, we can create “Create Table” script based on this
variable, and then store the script into a SSIS package variable. After that, we create a staging table based on the package variable, load the flat file data to the staging table. Eventually, we load data from the staging table to the destination table. For
the detail steps, please walk through the following blog:
http://www.citagus.com/citagus/blog/importing-from-flat-file-with-dynamic-columns/
Regards,
Mike Yin
TechNet Community Support

Installing Multiple Operating Systems with grub and Arch Linux

NOTE: Please keep in mind that there are many different ways to achieve this same result using various loop and ramdisk methods, read this with a separate window to jot down your comments and suggestions... this is ongoing for me so any help would be appreciated!
Read the full article at Install Multiple Os without cds
This is my first post and I plan on making this topic an official HOWTO with www.tldp.org.
I have been into the computer security scene since 1990, but I realized that I had very little experience with the various LInux, Unix, and alternative Operating systems out there.
I have a CD-RW drive but being a struggling computer security researcher I had no money for blank cd-recordables. What follows is how I managed to install various operating systems on my computer (1 hard drive) without having to burn to a CD the ISO and then boot from that.
I first partitioned my 120GB harddrive into 10 partitions, the 2nd partition is a small swap and the last partition is extra large because it holds all the ISO images..
I then wrote a small shell script to automatically download (I love wget!) the following.
OpenBSD
IpCOP
Libranet
Arch-Linux
Fire
Local Area Security
Packet Master
Devil-Linux
FreeBSD
Knoppix
Helix
Gentoo
Yoper-Linux
NetBSD
RedHat
Slackware
The script also downloaded Installation manuals and md5 checksums.. (let me know if I should post... its pretty unsophisticated
I installed Slackware (personal favorite) on hda1 using my last blank CD-R, note that I do not have a separate boot partitino. (Should I?). I also installed grub on the MBR. I love grub, if you read through the man pages and all info you can find about grub, you can learn a whole lot. Grub has much more features and capability than lilo, even though lilo comes installed by default with slack.
I organize my kernel situation as follows... In my /boot directory, I mkdir KERNEL, CONFIG, MAP, INITRD and that is a good way for me to keep my kernels and everything organized.. Another good way is a separate dir for each new kernel.
Since Arch-Linux is a solid distro, I'll use that as a first example.
Here is the Arch-Linux section of my shell script
goge Arch-Linux
$w http://puzzle.dl.sourceforge.net/sourceforge/archlinux/arch-0.6.iso
$w http://unc.dl.sourceforge.net/sourceforge/archlinux/arch-0.6.md5sum
$w http://www.archlinux.org/docs/en/guide/install/arch-install-guide.html
md55
cat arch-0.6.md5sum
md5sum arch-0.6.iso
md55
The first thing to do is to mount the downloaded ISO image so we can use it as if it were an actual CD.
mount -t iso9660 -o ro,loop=/dev/loop0 cdimage /mnt/cdrom
Where cdimage= the ISO image. EX. /usr/local/src/ISO/Linux/Arch-Linux/arch-0.6.iso
This mounts the iso as /mnt/cdrom.
Next you need to copy /mnt/cdrom to a separate partition for the booting process. So mkfs.ext2 /dev/hda9. ( I prefer reiserfs or even XFS to ext but if you use something other than ext2 you could run into some problems because some of the installation kernels and initrds don't include support for reiserfs and so can't recognize the files. Although you could use mkinitrd to create a new initrd with reiserfs support, that might be pushin it IMO... I use the 9th partition consistently for this. I know there is a "right" way to copy the /mnt/cdrom files so everything stays the way it is supposed too, using tar or cpio, but I'm lazy so I just do cp -rp.
(What is the tar or cpio commands to copy with correct permissions etc??)
So you mount the 9th partition as whatever, say /mnt/hd and then copy the files. Now what?
Now edit your /boot/grub/menu.lst file to include the specific options to boot arch-linux installation.
A good idea is to find the isolinux.cfg file somewhere on the distro cd, this will tell you what to include in the menu.lst.
Here is the section in my menu.lst
title Arch Install
root (hd0,8)
kernel /isolinux/vmlinuz load_ramdisk=1 prompt_ramdisk=0 root=/dev/rd/0
initrd=/isolinux/initrd.img
This should be self-explanatory. The root (hd0,8) is pointing to partition 9. So the rest of the commands start from partition 9.
When you experience problems, remember you can always edit the grub boot options by typing 'e' and then edit the section. Also, a good idea is to include several variations in your menu.lst so you can easily try other ways to boot efficiently. And, remember to read up on all the installation guides that come with your distro, specifically, hard-disk installs.
There are special cases, Gentoo, has a semi-new compressed filesystem called squashfs. BTW, this is AWESOME, so check it out. It has to be compiled into the kernel, so some work is in order, but use this recompile to optimize your kernel. You can get the squashfs patch for almost any kernel. I use the latest stable 2.6 kernel. Squashfs is incredible and although I don't think you need it to install from ISO, you do need it to expand the livecd.squashfs filesystem that comes with the cd.
Heres a sample Gentoo section from my menu.lst
title Gentoo Install
root (hd0,8)
kernel /isolinux/gentoo root=/dev/ram0
initrd=/isolinux/gentoo.igz init=/linuxrc acpi=off looptype=squashfs loop=/livecd.squashfs cdroot vga=791 splash=silent
A nother' tip is the shell that is provided if you experience problems, typically busybox or ash. The key tools to get you going from here is mount and chroot. Sometimes you will need to manually create a simulated file system and then chroot into it. For instance, you might have to create boot, etc, bin, directories on the target partition.
I generally install each OS onto the next partition (careful of the logical partition) and add it to my menu.lst after install. A good idea is after installation, copy the kernel and initrd(if there is one) to the slackware(or whatever) boot partition on hda1. I copy kernels to /boot/KERNEL/ and initrd's to /boot/INITRD, then menu.lst is more organized...
You then need to add an updated section to your menu.lst (just comment out the install section for later)
Here is the finished arch-linux section from menu.lst
title Arch Linux 6
root (hd0,2)
kernel /boot/vmlinuz26 ro root=/dev/hdc3
This doesn't use my convenient boot/KERNEL/vmlinuz26 as you can tell by setting the root to partition 3.
***NOTE: Make a backup of MBR using dd and save to floppy, also backup the partition table to floppy, using cfdisk or parted. And boot disks (I use 1 with grub, and 1 with slack, and tomsbootdisk) will invariably come in handy. Tomsbootdisk is recommended, and make the grub boot disk when you install grub. install to floppy.
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
The final result after some fun experimenting, is when I boot, I have a cool grub boot screen come up with the option to boot into whatever OS I want, this is handy for multiple reasons. One good thing to do after this is to port scan and vuln scan each OS, after you update of course. Write this stuff down and you will know the weaknesses/strengths of the various OS's.
I can boot a custom Firewall, snort, or multiple honeypots using this procedure, as well as a graphical kde environment with a kernel optimized for graphics and my processor/architecture, or an environment devoted to forensics or even an environment suitable for programming.
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
P.S. Some of the cooler alternative operating systems are BeOS 5, EOS, ER_OS, V2_OS, and my personal favorite Menuet. Menuet is 100% assembly graphical operating system that fits on a floppy. Its f'in money!
This should be a good enough example to get you started, this kind of thing should be learned and not just copied... Knowing how to do this stuff could prove to be exceptionally useful...

Start by reading all the articles built-in on your Mac - Help > Mac Help, search "printer sharing."
http://desk.stinkpot.org:8080/tricks/index.php/2008/04/how-to-print-to-a-cups-se rver-from-mac-os-x/
http://www.macosxhints.com/article.php?story=20080324224027152&query=share%2Bpri nter
http://members.cox.net/18james/osxprintersharing.html
http://ubuntuforums.org/archive/index.php/t-56940.html

IOS 8 creating multiple group messages with same recipients

After updating to iOS 8.0, my iPhone 5s has started creating multiple group messages with the same recipients. For example, my phone has created three group threads with the exact same people in it. Different friends' messages go to different threads. It is terrible confusing and annoying. It has done this with several of my group messages. Deleting the duplicate threads have not worked. They keep recreating themselves when certain people text the group.
Any suggestions? This has been a problem for many of my friends as well.

Hi hkapn,
Do you and your wife use the same Apple ID and are connected to the same wifi network? Your phone ringing to multiple devices is called Continuity and is a feature of iOS8. https://www.apple.com/ios/whats-new/continuity/
Essentially, it allows you to use your iPhone cellular connection on all of your iOS8 devices: iPhone, iPad, Mac computers with Yosemite. On some devices the call will look like it is coming through FaceTime, but likely on your wife's iPhone it just looked like a regular call.
Since your iPhone was "hosting" the phone call through it's own cellular connection, you were able to pick it up after your wife answered, which caused your wife's phone to disconnect. This is called Handoff and is also a feature of iOS8.
You can turn off Continuity if you'd like, but you could also look into getting your wife her own Apple ID. I'm guessing you were sharing one so that you didn't have to make purchases twice. Also in iOS8, there is a feature called Family Sharing which allows you to link multiple Apple IDs and share purchases (among other things).
In regards to the group text message issue, as far as I can tell, splitting group messages is NOT a feature. I agree that there should be an answer to this already!

Multiple-reader environments with shared filesystem?

Similar Messages

Maybe you are looking for